Aspects of the history of English pronunciation in Scotland

Abstract

The aim of this thesis has been to review, in the light of modern linguistic theory, a hypothesis put forward hy J.A.H. Murray in his .monograph "The Dialect of the Southern Counties of Scotland. London, 1873", according to which the confusion of the spellings <a> and <ai>, <e> and <ei> , <o> and <oi>, <u> and <ui> was due to a Scottish sound change in the middle of the 15th century resulting in a monophthongisation of the former /i/- diphthongs and a subsequent coalescence with the corresponding monophthongs.In Part I, I have examined the categories that guide work in "descriptive" and in "historical" linguistics and have elaborated a new synthesis of the two fields based on the categories level, unit, structure, system, dialect, and period. They are nothing but useful mental constructs invented by the linguist to make systematic statements about human language, but have no existence. Similarly sound laws are only formulaic expressions of correspondences, they do not represent the actual change.I have also maintained that historical (more precisely comparative genetic) studies are built on descriptive ones and conversely that they contribute to the latter by widening our knowledge of the spoken medium of past stages of languages. This is in fact their only purpose. Evolutionary formulae are an extremely powerful tool for reconstructing phonological features, which can be gained only in a very imperfect and rudimentary way from written evidence. In performing these operations, synchronic genetic groupings relying on certain regular correspondences on the phonological and formal levels of contemporaneous dialects are taken to reflect diachronic genetic states. This shift is made possible "by written documents, which act as a guide in reconstruction without "being able by themselves to give comprehensive information about the spoken medium. When reconstruction is achieved and documentary evidence supplemented by phonological data the comparative genetic constructs are eliminated.Modern dialects are thus of prime importance in evolutionary studies and provide the only safe way to establish phonological items of past stages; spellings are of minor value in this respect. They contribute to the solution of our task by making language form of the past accessible. Our reconstruction is thus only partial and not an entire projection of concocted, forms into the void. As the informational value of the two mediums is so different we should use them as separate levels of analysis.Part II is a reconsideration of Murray's hypothesis in the light of this theoretical discussion. His theory violated the above-mentioned methodological principle in that he relied partly on spellings, partly on modern dialects - depending on his familiarity with them in the different areas - and treated them as equal sources of evidence. A detailed study has shown that this procedure resulted in a false hypothesis. The modern data do not corroborate his view. In most dialects the items /a:/ and /ai/ are still distinct, and even in those areas where this is not the case today we have been able to demonstrate that the convergence happened after the Middle Scottish period, /o:/ and /oi/ are differentiated in all the modern dialects. The Middle Scottish rhymes quoted by Murray in support of this theory can all be accounted for by the same formulae as have been set up for the present-day material. A thorough investigation of the promiscuous use of the graphemes <a>, <ai> ; <e>, <ei> ; <o>, <oi>; <u>, <ui> has finally been found to be a scribal practice, which originated under special cultural and linguistic conditions, different for each of the four pairs.The results of the thesis are thus a contribution to the knowledge of the spoken medium of Middle Scots gained with the help of fictional evolutionary formulae, and a new interpretation of the written aspect of this ' etat de langue'. It can no longer be doubted in view of this that a handling of the two media in their own right is not only a possibility, but a methodological necessity to avoid serious errors in our linguistic analyses.